Welcome to Podcast Notes!
This feature summarizes a podcast about an interesting XR topic. Following up on last week's podcast about brain-computer interfaces and AR, our latest discussion features Elon Musk.
The Tesla CEO spoke on the Lex Fridman AI Podcast about the engineering and innovation at Neuralink, his company developing brain-computer interfaces, which could be fused with AR/VR applications for highly realistic immersive scenarios.
A quick recap: BCIs connect our brains to machines. Experts believe they'll be the driving force behind neuroreality, where we'll be able to navigate virtual worlds simply by thinking.
Companies like EyeMynd and Neurable are already developing systems that use this "brainwave control," which will bypass our senses to move around virtually. Facebook — which recently showed off its wristband for scanning electrical signals to predict what a user wants to do in AR/VR — also wants to build direct interfaces for interacting in VR using nothing but your mind.
Read on to find out what Neuralink is doing. [Note: This episode was edited and condensed for clarity.]
Lex Fridman: The work at Neuralink promises to help treat neurobiological diseases and help humans further understand the connection between individual neurons and the high-level function of the human brain. One day, it could expand the brain's capacity through two-way communication with computational devices, the internet, and artificial intelligence systems.
There is a lot of fascinating innovation across different disciplines going on at Neuralink. The flexible wires, the robotics sewing machine, the response to brain movement, and so on. Do you hope that the work at Neuralink will help us understand more about the human mind about the brain?
Musk: Yeah, I think the work in Neuralink will definitely shed a lot of insight into how the brain works. Right now, the data we have regarding how the brain works is very limited. We've got fMRI (functional magnetic resonance imaging), which is like putting a stethoscope on the outside of a factory wall and you can hear the sounds, but you don't know what the machines are doing. You can infer a few things, but it's a very broad brushstroke.
In order to really know what's going on in the brain, you have to have high-precision sensors and stimulus and response. If you trigger a neuron, how do you feel? How does it change your perception of the world?
It's having high-precision sensors that tell you what individual neurons are doing and then triggering a neuron to see what the response is in the brain, so you can see the consequences. It'll be really profound to have this in people because they can articulate their changes — for example, if there's a change in mood, or they can see better or hear better, or be able to form sentences better or worse.
Fridman: On the human side, there's some plasticity in the brain, and on the machine side, we have neural networks, machine learning, and artificial intelligence. It's able to adjust and figure out signals. So there's a mysterious language that we don't perfectly understand that's within the human brain. And then we're trying to understand that language to communicate both directions. Where do you see the biggest benefit arriving from: on the machine side or the human side? Do you see them working together?
Musk: I think the machine side is far more malleable than the biological side by a huge amount. It will be the machine that adapts to the brain, which is the only thing that's possible. The brain can't adapt that well to the machine. You can't have neurons start to regard an electrode as another neuron.
We currently operate on two layers. We have a limbic primitive brain layer, which is where all of our impulses are coming from. It's like we've got a monkey brain with a computer stuck on it, which is the human brain. And a lot of our impulses are driven by the monkey brain, while the computer is constantly trying to make the monkey brain happy. It's not the cortex that's steering the monkey brain; the monkey brain's steering the cortex. The cortex is what we call human intelligence, like an advanced computer, relative to other creatures.
Fridman: How will machine learning methods, like natural language processing applications, be applied for communication between the machine and the brain? Do you see the value?
Musk: Yeah, absolutely. We're a neural net and AI is basically a neural net. So it's like a digital neural net will interface with a biological neural net and hopefully bring us along for the ride.
The vast majority of our intelligence will be digital. Think of the difference in intelligence between your cortex and your limbic system, which is gigantic. Your limbic system really has no comprehension of what the cortex is doing.
Fridman: What is the most exciting you see for the future impact of Neuralink, both on the science and societal impact?
Musk: Neuralink will at first solve a lot of brain-related diseases, like autism, schizophrenia, and memory loss. There's a tremendous amount of good that Neuralink can do in solving critical damage to the brain or the spinal cord. There's a lot that can be done to improve the quality of life of individuals.
Giving somebody back full motor control after they've had a spinal cord injury, restoring brain functionality after a stroke, and solving debilitating genetic brain diseases are all incredibly great. In order to do these, you have to be able to interface with the neurons at a detailed level. You need to be able to fire and read the right neurons. And then you create a circuit, replace what's broken with silicon, and essentially fill in the missing functionality.
Over time, we can develop a tertiary layer. The limbic system is a primary layer and the cortex is a second layer, which is vastly more intelligent. The tertiary layer will be digital superintelligence. It will be vastly more intelligent than the cortex but still coexist peacefully in a benign manner.
Ultimately, though, it's intended to address the rest of the existential risk associated with digital superintelligence. We will not be able to be smarter than a digital supercomputer. So, therefore, if you cannot beat them, join them.
It's important that Neuralink solves this problem sooner rather than later because the point at which we have digital superintelligence, things become very uncertain. It doesn't mean that they're necessarily bad or good, but passing the singularity, they become extremely unstable. We want a human brain interface before the singularity, or not long after it, to minimize existential risk for humanity and consciousness.
The problems at Neuralink are material science, electrical and mechanical engineering, software, and microfabrication. What it comes down to is you have to have a tiny electrode that's so small, it doesn't hurt neurons. But it's got to last for as long as a person. You've got to take that signal and process it locally at low power. We need chip design engineers because we need to do signal processing in a very power-efficient way. The brain's very heat sensitive.
Then we need to automate the whole thing [with robots]. It's like LASIK. If done by neurosurgeons, there's no way it can scale to large numbers of people. And it needs to scale to many people, as ultimately we want the future to be determined by a large number of humans.